The State of HAB Monitoring – Session Two Innovations in HAB Monitoring Technology

September 24, 2024

Image of a marina at sunset with

Session Two: "Innovations in HAB Monitoring Technology" Managing Harmful Algal Blooms (HABs) has been a priority across the globe for years and the 2014 Toledo Water Crisis elevated this challenge to become one of the most pressing Great Lakes water quality concerns. Lab analysis and optical sensors have long been the go-to methods for monitoring HABs, but the emergence of new solutions leveraging AI-powered microscopy, eDNA and eRNA, and low cost sensors may revolutionize our approach to collecting these critical data. This panel discussion will engage industry experts and research leaders to explore the next generation of HAB monitoring technology.

About the speakers

Igor Mrdjen

Program and Science Lead
at
Bloom Optix

Dr. Igor Mrdjen is the founder and Science Lead of BloomOptix, LLC, a start-up using AI and portable microscopes to empower more users in their HAB monitoring needs. Igor is based out of Syracuse, NY and since his days at Ohio State University has been studying the public health and environmental impacts of algal blooms. For over a decade he has strived to expand our understanding and capabilities in reducing algal bloom exposures and occurrences.

John Higley

CEO & Principal Scientist
at
EQO, Inc.

John started research for what became EQO in 2016 with the goal of bringing the modern molecular biology lab to the field of environmental biology. Prior to forming EQO, John worked on early cancer diagnostic platforms and microbiome analytics at Thermo Fisher Scientific and led the experimental therapeutics department as the Senior Director of R&D at Molecular Templates. John founded EQO following the successful development of EQO’s flagship detection platform; originally developed for a grant-funded study of three Texas lakes. At EQO, John has further developed RNA preservation and population dynamic analytics systems, a patent pending automated collection and preservation platform, and is developing an NSF and US-BoR funded biopesticide for mitigation and eradication of zebra mussel infestation. John is a science advisor to the interagency working group on eDNA, which includes representatives from USGS (United States Geological Survey), US-BoR (United States Bureau of Reclamation), USFWS (U.S. Fish and Wildlife Service), and multiple state Departments of Natural Resources.

Chris Lee

Co-Founder and CEO
at
AquaRealTime

Christopher Lee has 20 years experience bringing to market full stack technology products, the last 10 in the water monitoring space. He has a PhD in Electrical Engineering. In his spare time, Chris is an avid runner and skier and loves exploring the backcountry.

Thomas Bridgeman

Director of the Lake Erie Center
at
University of Toledo

Dr. Thomas Bridgeman is a Professor in the Department of Environmental Sciences at the University of Toledo and the Director of the UToledo Lake Erie Center. His research centers on environmental challenges facing Lake Erie, especially harmful algal blooms (HABs) and ‘dead zones.’  His research includes improved bloom monitoring technology for drinking water treatment plants and the influence of episodic hypoxia, river plumes, and lake sediments on the growth of HABs. Tom teaches courses in aquatic ecology and currently serves as an advisor to ODHE’s Harmful Algal Bloom Research Initiative (HABRI) and as past president of the Lake Erie Area Research Network (LEARN) He earned degrees from Miami University, The Ohio State University (MS), and the University of Michigan (PhD).

[00:00:00] Ryan Sorichetti: All right. Good afternoon, everyone, and happy lunch hour for everyone in the Eastern time zone, my name is Ryan Sorichetti, and I'm going to be the facilitator and moderator of our discussion today. here, we're going to be focusing on the state of HABs monitoring with a particular focus on innovations in HAB monitoring technologies.

[00:00:34],this discussion today is part of a broader webinar series that's presented by the Cleveland Water Alliance, as well as the Great Lakes Water Authority. Commission's Harmful Algal Blooms Collaborative, and we're all here with a vested interest or experience in some capacity that identifies and recognizes the importance of HABs monitoring.

[00:00:57] Much of this is made possible through collaborations and partnerships, and you're going to hear the word collaboration quite a bit today.  just in terms of the Great Lakes Commission's collaborative, this was started in 2015 to really get at that. and bridge the communication gap that may exist between scientists, decision makers, policy makers regarding HABs and research that is conducted in the Great Lakes.

[00:01:21] This is done through multiple forums and venues, including conference sessions, like those that are hosted at the International Association for Great Lakes Research, various newsletters and publications, webinars. Such as this.  but the fabric of this is that the collaborative really works to improve communication through a common agendaddressing the science and management needs for that region.

[00:01:43] And that's why we are all here today. As mentioned, my name is Ryan Sorichetti and I'm a Great Lakes research scientist with the Ontario Ministry of Environment, Conservation and Parks. I'm a Great Lakes ecologist and my experience and background in the HABS world really spans from monitoring near shore areas for algal populations within the Great Lakes.

[00:02:10] And then our office and group also undertakes various research studies to get at more mechanistic and process based questions that are related to harmful algal blooms. I am the new Canadian co chair for the Great Lakes Commission's HABs Collaborative, and I'm very excited to be in this role and bring some of my knowledge and expertise to the table in that capacity.

[00:02:30] And I'm also very much looking forward to working with Janan Karbash at the University of Michigan, who's my complimentary U. S. co chair, and through this, we'll be working with the Habs Collaborative over the next 2 years in various forms and formats. But today we're here to have an open discussion in a broad discussion with some panelists or experts in the fields of HABs monitoring, particularly with those related to various technologies.

[00:02:58] And if you're not sitting in front of a computer, and you can't see the image that we have on the screen right now, I'll mention that those panelists are Igor Mrdjen,John Higley, Chris Lee and Tom Bridgman. I'm going to call on each of you individually in just a few moments to speak to some of your experiences and introduce yourself.

[00:03:17] But I  first, just wanted to highlight that throughout this discussion, you can submit questions through the Q&A  box that you can find at the bottom of your zoom screen. If you just bring that up and type your. Question into the box, I'll be monitoring this frequently throughout the discussion. And then at the end of the, at the end of our time here today, with 10 or 15 minutes left in the session, I will move everything in towards a question and answer period.

[00:03:43] And I'll be going to this chat feature here in the  Q&A box to bring those questions forward to our panelists. If there are any issues with that, just send me a message on the side and we can see if we can get you sorted. And just before we get started, and if you're not going to be here towards the end of the presentation, when I plan to pull this information slide back up again, as I did mention, this is part of a broader webinar series that's held by the Cleveland Water Alliance and the Great Lakes HABS Collaborative.

[00:04:13] The next one will be on Tuesday, October 17th, and the focus of that Will be on innovations in nutrient monitoring technologies.  and then,the, QR code is there that you can scan to either register. If you registered for this session, you may get information emails that will allow you to register through there.

[00:04:31] And there's also the website that is presented at the bottom there. If you can copy that down and join through there. So, what I may do at this point is stop sharing my screen, and we are going to open the discussion. And the  first thing I'd like to do is I'm going to call on each of the panelists individually.

[00:04:51] And I'd like you to just introduce yourselves, provide a little bit of background on your experiences working with HABs and monitoring and research technologies. And it'd be really helpful if you could frame some of that discussion in the scope of your work, scope your work in the context of HAB monitoring technologies and what values you are seeing of those technologies to date.

[00:05:12] So,  first, I'd like to call on Igor.

[00:05:16] Igor Mrjden: All right, I'll try to cover everything Ryan, but if I forget to cover something, just yell at me. So, like Ryan said, my name is Igor Mergin. I'm the founder of Bloom Optics. I've been working in HABs,since my PhD days at Ohio State University. And I was always interested in kind of how do we kind of address the issue of harmful algal blooms in, you know, in lakes that are often overlooked.

[00:05:46] In ponds as well, and environments that don't necessarily get a ton of grant funding, for example, places outside of Lake Erie. And how do we really make monitoring and diagnostic tools and mitigation efforts just kind of accessible to everybody? So.  I started working,  using just the same old regular stuff that everybody uses,you know, Secchi disks,things like that.

[00:06:16]  all of that,all of that kind of field equipment did a lot of genetic analyses, and saw really the power of that, of those technologies as well. but then I, you know, we quickly realized that we needed something. In the middle there, between those 2,  you know, the super, super rudimentary and the, like, Uber nerdy stuff,that, you know, grad students,geek out on,  and so,yeah, that eventually led to,  to the founding of, of Plume Optics with a mission of kind of,  empowering,average everyday people with,  better capabilities.

[00:06:53] So,

[00:06:56] Ryan Sorichetti: thanks Igor.,John.

[00:07:00] John Higley: Thanks,  kind of build off of what Igor just mentioned. I'll talk about us. So we do that nerdy stuff.  the high, the deep molecular kind of technology. My background is as a molecular biologist. EQO is a biotech company,  and we do, deep molecular analytics on nucleic acids.

[00:07:20] So DNA and RNA. Um. We've done HABs monitoring from both a perspective of, okay, what is there? What, what are these cyanobacteria that are living there? Then also, what are they expressing? How much of this toxin that we're concerned about are they expressed? Then also looking at the whole community, we do whole microbiome analytics to figure out what is the family?

[00:07:43] What are the other bacteria that are in the environment? How might they be interacting with each other? And. This is really kind of,zone deep into, like, impact. What's the impact on human  health? Where is it likely to go in the future? Are some of these relationships between bacterial populations?

[00:08:01] Can that allow us to predict?  those are the kind of questions that we answer, and that's the kind of work that we've done a lot around toxin expression and population dynamics of cyanobacteria.

[00:08:14] Ryan Sorichetti: Excellent. Thank you, John.,Tom.

[00:08:18] Thomas Bridgeman: Yeah, thanks.  I've been working on Western Lake Erie, you know, 1 of the hot spots for perhaps over 20 years now and seen development of a lot of techniques,  you know, early on. And when I was a grad student, we're mostly just looking through microscopes to identify what kind of value you were there and what kind of cyanobacteria were in the water.

[00:08:39] And then we had the early, the early fluorometers that could,identify.,total chlorophyll, and then fluorometers came along that could identify total chlorophyll plus phycocyanin, which is the pigment in blue green algae. And now there'sfluorometers that can identify different algal groups, you know, diatoms, green algae, blue green algae, and so on.

[00:09:01] And most recently,fluorometers that can do all that, but also add a physiological, component that can,measure the health of the cyanobacteria. So I don't develop those things, but I use them and I kind of talk with the developers about, you know, things that would would help development. So,that's in the detection of cyanobacteria.

[00:09:23] And then in the detection of toxins,  worked with, you know, everything from the early ELISA kits,  now,the development of faster ELISA sort of handheld ELISA kits, and then watching. From afar, the development of deployable toxin measurements,the ESPs that Noah and others are developing that can be tethered in the lake and,automatically take a sample and analyze it and send the results up on the web.

[00:09:51] So, so that's all things that I've seen,and, and worked with over time.

[00:09:58] Ryan Sorichetti: All right. Thank you, Tom.,lastly, Chris.

[00:10:02] Chris Lee: Yeah,thanks Ryan. So,  yeah, I'm Chris Lee. We,my co founder and I previously started another company that was on the other side of the coin. So we were making monitoring systems for commercial algae growers.

[00:10:16] So people who are making algae on purpose and in ponds and reactors, and then we realized, okay, there's, you know, there's this huge problem with.  there's so many locations that are dealing with harmful algal blooms, and people are spending a lot of time going out there and sampling. And the other option is to use some of the sensor systems, which can be extremely expensive.

[00:10:41] So that negates the ability for the average person or group to use them. So, we saw a real need for a device that would be buoy based and dashboard based. That would be very simple, but you could deploy to a lot of locations and critically, it would need to be low maintenance because and low cost. the low cost is, of course, important, but the real cost is over time.

[00:11:09] How much time do you spend with these devices? So, we saw a gap in the market for devices that would help you cover a lot of terrain, a lot of a lot of waterways. So, we had, we produced a device that that can be installed in 20 minutes. It's buoy based. Everything is pre assembled.,within 5 minutes, you can see it on the dashboard with.

[00:11:32] A learning systems and things like that. So this is sort of the. The

[00:11:40] more course of the technologies in this group, but intended to cover that  first, line of defense out there in the field. And then after that, let's bring in more species identification with ecosystem or bacterial colonies, different things like that that are more in depth.

[00:11:57] Ryan Sorichetti: All right, very interesting. Thank you all for those introductions.

[00:11:59] So the common thread that I see running through this group is we're really all working towards preventing negative impacts to human health and the environment by way of exposure and occurrences of HABs, and we've come a long way. So, traditionally, and historically, it's been looking through microscopes, physically being there when a harmful algal bloom occurs.

[00:12:19] So you have to be there to see it or be out there prior during the precondition. And it sounds like everybody's working a little bit towards,being in more places,  for longer periods of time to really enhancing that spatial and temporal time scale that we can have the ability to sort of monitor research and respond to harmful algal blooms.

[00:12:42] So, that's all very interesting. So thank you again for those. What I'd like to turn the discussion to at this point is some of the challenges and current, limitations that, Impacts of emerging,  have monitoring technologies have. So maybe I'd like to hear some about some information about current adoption gaps, implementation barriers for the new technologies that you are working to implement.

[00:13:08] Or some technical hurdles that are impeding the scope of progress of data driven have management.  so, for example, John,  I know that the field of DNA and RNA is rapidly evolving. It's I feel that every time we.  go to international conferences related to HABs. There's new developments, emerging developments, but it is a very fast moving field.

[00:13:33]  I was wondering if you could weigh in on some of that related to some of the hindrances or gaps that we might be experiencing still presently.

[00:13:41] John Higley: Yeah, sure. I mean, so 1 with molecular tools, just like any of the other tools that we've been talking about today, and that we will be talking about today. We have to look at all these as a suite of tools.

[00:13:51] What are these individual tools good at? What are they not great at? And understand where they play their role. So, when it comes to molecular diagnostics,  some of the hurdles are really just around education. What is available? What does it actually mean? What does it tell you? So DNA can tell you some things.

[00:14:10] RNA can tell you a lot more. That's kind of where we specialize.  and one of the big issues are, I hate to say it this way, but people over selling certain things. So NGS is very popular right now.  and it stands for next generation sequencing. We see a ton of people doing next generation sequencing.

[00:14:31] Next generation sequencing certainly has its place. However, it is not quantitative. Every time you run an NGS, you got a different denominator. So you don't have reproducibility. So you'll get to run a test that's very expensive and you get. Percentages and those percentages are going to change whenever you do it again, because the denominator changes  every time just because it's a, I won't get into all the chemistry as to why that is, but it's,an issue.

[00:14:57] What it does tell you is a great example of a snapshot of what everything is that's there and you can then also do that on the transcriptomic side of what is everybody producing. Right? These are great possibilities. These are things that are there, but there are a lot of times where we see NGS utilized where it should not be utilized and we see QPCR utilized for maybe the wrong thing.

[00:15:21] QPCR is very quantitative. It's very directed, but it's after like 1 thing. I know the thing I'm looking for, and I want to quantitate that. So we have to understand how these tools work to be able to understand how to best implement them. And there's a educational issue that's largely due to, I would say, unscrupulous marketing, which is a little bit rough to say in my field, as I'm one of those people who markets molecular tools and molecular diagnostic tools, but it is a reality.

[00:15:50] And additionally, you know, one of the big hindrances to molecular diagnostics is that we have to understand that it has to be a combinatorial therapeutic. We all look at these,You know, I'm sure we've all seen nature documentaries or like Cosmos or planet Earth or whatever. You always have that big, you know, zoom out scale of here's the earth.

[00:16:09] And then we zoom down and we zoom down. We zoom down and then we get to this little mouse or whatever you can zoom down further. You can go down to the atoms. Right? So, you know, we're looking at the molecular scale of this area of that point. We need to understand the scope of what that is, what that means and how it translates to the larger scale.

[00:16:27] And in that process of zooming in. Okay. Satellite imagery is important. Fluorometers are important. Protein, you know, ELISAs, there are, I can get into the technical issues of where ELISAs work really well and where there's some issues, you know, there's so many technologies along the way. I think scientifically, we're at a moment where we have so many amazing tools.

[00:16:50] But finding a kind of a cohesive plan of attack is an issue. And,I think that honestly is the largest issue because there's a ton of funding in this area. There's a ton of,people who care about it. It obviously has huge impact to a huge amount of people.

[00:17:12]  a lot of it has to do with figuring out a unified theory of attack and understanding, and a large scope, like, educational scope of what these tools are, what the limitations are, what they do, what they can't do, and then ultimately from the detection side. Diagnostics is great. It should go a step further and we should also go to okay.

[00:17:36] Great. So we know where the problem is going to be. We know what it's going to do. We can minimize impact. Where's it coming from? How do we stop it from happening? Right? So, you know, the reason why, when you go to the doctor, you have a cancer scare, the diagnostics also ultimately need to lead to a treatment.

[00:17:54] And that's something that I think we have the tools for. It's just finding that unified theory and collaboration. very much.

[00:18:05] Ryan Sorichetti: Very interesting. Thanks for that just in terms of the limitations on some of those tools, where does the state of the science lay right now in terms of the reference information that we have to compare our environmental samples with what's known in terms of eDNA and eRNA profiles?

[00:18:20] John Higley: That's phenomenal.

[00:18:21] So,  NCBI was originally put together for research for human health, but basically anytime anybody sequences anything, it goes into that database and everybody has access to it. A lot of companies saying we have this database. It's NCBI. Everybody's using BLAST. Right? So it's a great database. Like the depth of knowledge is huge.

[00:18:43] We have sequenced and. Yeah. And have very good understanding of functional aspects of over 50 micro system toxins at this point,  as well as. And, you know, more endosomal toxins, like saxitoxins and anatoxins that aren't released until the, until the bacteria until the cyanobacteria lysis, we have very good depth of knowledge on what these things are and what they do.

[00:19:11] We also have an understanding of the different operon families of when they're turned on, when they're turned off, what causes them to turn on and off. I mean, there's always more to learn, but there's pretty good depth of knowledge in this field.

[00:19:25] Ryan Sorichetti: Thank you. That's great. And if anyone else has questions for John, please feel free to drop them in the Q&A section.

[00:19:31] I'll bring those back up towards the end.  at this point, I'd like to turn focus over to Chris a little bit, because I know that the technologies that you're working on are really striving to be.  in more places for long period, longer periods of time, you know, capturing the bloom, but also being there in the precondition, the post condition, and the work that you've done is really striving towards increasing our capacity to be a little closer towards everywhere at, you know, for a longer period of time in doing that.

[00:19:58] Are you finding any gaps in your ability to reach that goal?

[00:20:03] Chris Lee:  so, well, basically we are seeing, you know, quite a lot of the same things that John mentioned. So,  you mentioned,like,pre jury and post treatment. That's. That's kind of a slam dunk application where we can do really well, you can see high values in.

[00:20:23] Chlorophyll, phycocyanin, turbidity, you name it, you can identify that it's cyanobacteriand then you can see a dropping. So that's great.  the, what if it's things are kind of not that bad yet? That's when people are a little bit having trouble deciding when do I care? I'm seeing these cool plots in the dashboard.

[00:20:45] I'm seeing, you know, it's going up and down. Then where do I actually care? And where do I need to start taking measurements of species, of bacteria, of toxins? And then, you know, zooming out a bit more, if I get a species and I see that yes there's Microcystis,  well, now I want to know, is there the gene that might be producing a toxin?

[00:21:08] So then now  then we see that there's a gene now, I still don't know if there's a toxin. So that whole final answer is not there.  so the,  in fields like this, where there isn't that final slam dunk, quite often people go towards a so called evidence based approach where they say, okay, we don't have the perfect answer for you.

[00:21:33] But we have seen that when, for example, I'm just going to pull numbers out here when the chlorophyll a is above 10 micrograms per liter that that associates with a higher percent, maybe a 70 percent chance of producing a toxin eventually. And people have done that kind of work. I, things that come to mind is the government of Canada has some had published something last year.

[00:22:00] The state of Kansas had a sort of a white paper. So, I think we're in a position to take,bit of a step back and, you know, this is sort of what John was alluding to a unified approach and say, look, we've got these. Values in this technology, we've got that value in this technology. We feel confident that we need to do something about the problem and spend a bit more money or effort or time on this.

[00:22:27] So, whereas we don't have the capability now to. Give the final answer, I think we are in a position to start making kind of these look up tables of evidence based approaches of procedures. Let's do this. When that happens, let's do that. When this happens. I think Tom would be a Probably a better one  to assess whether that's feasible at this time, but that's sort of an idea that I think we're striving towards.

[00:22:53] Ryan Sorichetti: Right, and I was going to kind of go there next with Tom, just to ask. I know that you're always, you know, striving and working with the latest technologies and some of the, some of the more intricate suites of equipment and monitoring technologies that we have available. And in working with that equipment, are you finding that there are hindrances in your ability to sort of deploy these in strategic areas, or use them effectively in hives, monitoring or research?

[00:23:21] Thomas Bridgeman: Yeah, sure. It's a, it's a matter of scale of,resources that you have versus questions you want to answer. It's sort of, I think, in terms of like, what's the information you need and then how fast do you need it? And then how much money do you have to spend?  And,here I work with a lot of well, I work with mainly with the Toledo drinking water plant that serves half a million people and,you know,they do have some money to spend.

[00:23:49] So they can, buy more advanced equipment. And they need to know they need to have an answer right away. They need to know things within hours, not days or weeks. So that's when, you know, some of the molecular tools take a lot of time and the fluorometry can be fast. So we do our fluorometry quite a bit.

[00:24:09] Um. But the downside is the instrument that we're using that it does a online algal interpretation about every 15 minutes puts it on the Web. That instrument's very expensive, you know, maybe 35, 45 thousand dollars , You can't buy a lot of those, but for a targeted application, like a big drinking water plant, you know,  that's the kind of place where those things go.

[00:24:33]  you know, otherwise you can use the usual sort of florometers, put them on buoys.  and get a lot of information, but not at the detail that a drinking water plant needs and, and not in at the, the same rate that, that, that they provide.  so, and, and the limitations, even the most advanced omet can't tell you anything about toxin.

[00:24:57] So that's where, what Chris was talking about, you use your past experience and, and,you start to say, well, if we see these levels of cyanobacteria or these levels of chlorophyll in the past, it has course. correlated to these levels of toxins. So you're not, you know, you can kind of infer from the past what you might have now.

[00:25:15]  and that's because, you know, toxin measurement still takes time. That's, that's one of the, the, that's one of the challenges we still have is that it still takes time. It still takes some fairly advanced instrumentation to directly measure toxins. So you kind of have to infer it from past experience.

[00:25:34] Ryan Sorichetti: Let's see, so it sounds like we're getting really good at collecting high volumes of data and information.  and Igor, I could imagine that. Your machine learning applications and models and what you're working with automation and those machine learning efforts would require a high volume of data in order for those to be effectively sort of implemented for your goals.

[00:25:57] Even though the state of the science is progressing to the point where we're really good at collecting high volumes of data and information. Are you finding that enough to develop the machine learning applications that you're intending?

[00:26:08] Igor Mrjden: Yeah, so that's a really complicated question because machine learning is a very complicated,field.

[00:26:15]  and but what we found is that, you know, you can use machine learning as a very good,  way to kind of, you know, Automate very simple, relatively speaking tasks, right? So, for us, it was basically taking 20, 000 or so microscopic images of cyanobacteria, feeding them into a machine learning model and helping the machine learning model basically learn how to classify and enumerate the cyanobacteria under the microscope.

[00:26:49] Right? So, like Tom mentioned, that used to be something that you would take a sample. Ship it off to a lab and somebody would look under a microscope and basically have a little counter and click on the counter and count for that's all they would do all day basically, because everybody from every lake is sending multiple samples off to that lab.

[00:27:12] Right? And there's huge backlog. So, by the time you sampled your lake. And sent it off to a lab, it would be a week or 2 before you heard back on whether or not you had a bloom and what was in it and what was causing it. Right? So it was really good for kind of like historic data keeping and kind of.

[00:27:30] You know, planning for next year's kind  of response. It was never really meant to be something that you would say, I'm going to take a sample, see what's in it. And then I'm going to act on that information immediately. Right? It was never that fast and that good enough. The beauty of it is, is that now that we're advancing so much in the hardware of, like, microscopy, right?

[00:27:49] Like, we work with a company, they make these, like, this is, this is our microscope, right? That's how tiny it is and it hooks up to your cell phone. Right. And so all of the computing power is on the Internet and on your cell phone through the app. And so you can equip basically anybody, whether it's a lifeguard on a beach or a citizen scientist, or an expert like Tom, and you can equip those people with things like that.

[00:28:13] With this ability to take photos in the field and then they can analyze them with the machine learning aspects. And the really nice thing is about machine learning is the more people you get using it, the more information you have in your library, the better it gets over time in terms of accuracy.  and so the real beauty of it is, is because you can do all of this through a cell phone.

[00:28:35] You can also add GPS locations to all of the data you collect and use your cell phone as the.  basically,  your field data sheet. So you can also add observations like, oh, I noticed that this beach was closed. Then an alarm was raised. I also did some nutrient analysis back at the lab and I correlated this data and then, but just by submitting that data, you know,for analysis anduploading it to the database, it can basically automated automatically be incorporated and combined with higher,instrument data, like Chris's buoy data or John's genetic data.

[00:29:14] And we can build these huge data sets with minimal effort, basically, that then can be handed off to to folks to use for predictive modeling,have forecasts, um.  Any of that kind of stuff, right? So  the future is really kind of how do we make these observations and these analyses more rapid and more accessible to where you don't have to have a Ph.

[00:29:41] D. to use these tools.  and where the results are meaningful, right? It's 1 thing to give somebody,a print out that says, hey. You have a have in your lake versus here's all of the data at your fingertips in graphs and photos in readouts that you can see basically in real time to tell you exactly what's going on this moment.

[00:30:07] And. On the back end of that, by making that much faster and much more accessible, you're increasing the number of data points you can gather on your lake. Right? Before we were constrained with how many samples can you collect? How many can you afford to ship to a lab? And how many can you afford to get analyzed?

[00:30:24] And how many can you afford to send off for genetic sequencing and all that kind of fun stuff? Right? By lowering the cost of that sample analysis and increasing the speed, you can cover more ground. You can decide strategically, which samples you want to know more about and you can basically get more people involved.

[00:30:42] So, you don't have to hire as much staff. You can use citizen science and volunteers to to handle a lot of your workloads, which ultimately gets to the point of kind of making it more possible to get the science out there. Outside of the folks with a ton of money to spend on expensive instrumentation.

[00:31:01] Ryan Sorichetti: Yeah, thanks for that. That's very interesting, especially factoring in ideas like citizen science and things to overcome these challenges and barriers and not to single out or point to any of the panelists individually. But does anyone else have any sort of low hanging fruit or ideas or things that could be implemented to overcome the barriers that you are currently experiencing or that you foresee?

[00:31:26] John Higley: Not low hanging fruit. I mean, the biggest thing has to be collaboration because the issue is that You know,  genetic data as Tom was mentioning, you know, it takes time, especially if you want to do something that's worthwhile. Like, there are ways to do DNA very quickly in the field. But it's not quantitative and doesn't tell you much, you know, and so really have to do a little bit deeper of analysis that requires sending it off to the lab.

[00:31:50] And so, since that's not real time, the value of that data comes into what else is it associated with? What other data do we have? Like, like it's an incredibly important piece. I, I, I hesitate to go into all the chemistry of, of all the different molecular tools, but the point is that it's, There's a lot of power there in determining specifically what toxins are being produced.

[00:32:15] You know what toxins are here best to send off,  to to get really clear answers on that best to send off to get really clear answers on what's been produced and to look at the total population from a semi quantitative perspective using like microarray or a non quantitative perspective using things like NGS, there's value in all these tools, but they take time and you're not going to get the answers right away.

[00:32:41] So you have to pair them with other data that is real time that is on that point to make to make them useful for things like, okay, what is this correlated with? Why are we seeing this? How can we prevent this from happening in the future? What was the impact or the potential impact of this event? That information only comes from it being, you know, kind of in a larger model.

[00:33:12] Chris Lee: Another thing that comes to mind is,  I think,in some cases, you can take that thinking of combining the things a step further. There's, there is potential for,putting, um. For example, species identification,  on a buoy that's low cost.  that's something that we've played with,as an idea and,we're, you know, there's different, um.

[00:33:44] Organizations out there that are providing grants. We have a NOAA grant recently.  and I do think that,some of this stuff. You know, imagine that you could have a buoy that was doing species ID. And then you created this sort of evidence based table that said, if you have this kind of algae and this concentrations.

[00:34:06] Now, we're starting to be a little bit more confident that this correlates with. A problem no, we don't know the toxins yet.  now, maybe, you know, going even further ahead, there will be something real time for that. I wouldn't be surprised if that happens at some point. Although I'm not the right person to ask that question too, but.

[00:34:27] I think combining technologies, some of them can be combined in the relatively short term talking years, not decades. And the other aspect that comes to mind is integration with treatment modalities. And I'm sure there'll be a panel that's going to talk about that maybe in more detail, but there's a lot of new modalities for treatments coming out in the past when, you know, if you ask someone, do you ever treat your link?

[00:35:01] Or your reservoir, the answer is, oh, we can't use copper. It's no, we don't do any treatments. Well, there's a lot more than copper coming up now. There's nano bubblers. There's encapsulated other products. There's ultrasonics, you know, the, the results at this point are mixed and sometimes, you know, we're not sure yet about them, but.

[00:35:23] It's clear that some of them are getting much better and are going to be very effective, including scale.  and so, as we get to that point, then we can start making really good overall decisions. Like, you've got this evidence based,  idea of of the likelihood of a problem, like a predictive risk index.

[00:35:46] And then you've got that matches well, for this location, you're not allowed to use copper. Here's another option for treatments, you know, getting getting to the point where we have a big picture, including the solution. And obviously, the real solution is prevention in the  first place, but that's yet another can of worms.

[00:36:06] So.

[00:36:09] John Higley: That's where we're trying to come in and understand those kind of things before it happens.

[00:36:15] Ryan Sorichetti: Yeah. Thanks, Chris. So, if anyone has any other questions for the panelists related to those topics, you can put them in the chat, but I'm going to ask the panelists to open your minds a little bit here and sort of envision what the future of HABs monitoring looks like.

[00:36:32] So, maybe we'll start with Tom. How do you envision the potential ultimate impact of the current and future of HABs monitoring technologies? Chris. I

[00:36:49] Thomas Bridgeman: sneezed earlier, so I had to mute myself.  yeah, I can see advancement along a lot of different fronts. I'm excited about the sorts of things that Igor is doing with the optical identification of,  of different phytoplankton species andkind of,you know. Artificial intelligence,you know, learn quantifying them and it feel like it's come a long way.

[00:37:15] The optical techniques were not so good until very recently, but I think they're advancing very quickly. And as Chris said, we might even be able to deploy those out on a buoy or something. That would be fantastic. I mean, other advancements are, I think we are making advancements in sort of the toxin detection.

[00:37:34] There are our kits now that are can be used by a citizen scientist and it's sort of a miniaturized ELISA on a chip sort of technique. They still have limitations though, in that.  in order to measure toxin directly, you have to break open the cells and release the toxin into the water so that it can react with the enzymes.

[00:37:57]  so there's that sort of freeze thaw sort of step that's involved in, in,ELISA toxin measurement is still kind of a barrier. You can't just go down to the beach and take a toxin measurement unless you, you know, unless you have a can of liquid nitrogen or something with you to do some freeze thaw.

[00:38:15]  so that there's. That needs to come along. And then,  at the high end, there's the kind of stuff that Noah is doing. And I see Greg do sets on the call, and he knows, you know, he's the expert on this, but they are miniaturizing the ELISA techniques are similar to ELISA and putting them on instruments that can sit on the bottom of the lake and take a measurement or even be put on a sort of torpedo type device and travel around the lakes, taking samples and performing ELISA measurements on them.

[00:38:44] And that's, and that's really. That's really interesting. Of course, that's very high end. We're talking millions of dollars at that point. And then,  1 of the things that's come about recently is, is a new interest in,the form that toxins are taking, especially are they in the cells? Or are they outside the cells?

[00:39:06] And,that's the real concern for people living along lake shores. If cells are breaking open.  that can lead to aerosolization of toxin. And so people on boats and people living along the lake shore can breathe in the toxins. Also a big concern for water treatment plants. If the toxins are outside the cells now, they are not getting filtered out.

[00:39:25] They're going right through the water treatment plant, and they have to be destroyed by adding more chlorine or ozone or something. So that's,fortunately, some of the instruments I've been working with are, you know, Have capability to detect when cyanobacteria cells are breaking open and in real time.

[00:39:44] So that's really useful. If you need to know that in real time, so all of those ways, I, see sort of the rapid detection of blooms and their toxins are advancing.

[00:39:58] Ryan Sorichetti: Great. Yeah, and I think that sort of parlays naturally into,some of Igor's work, right? So I was just wondering, where do you think the future of HABs monitoring is going with if our ability to improve on these technologies persists?

[00:40:13] Igor Mrjden: Yeah, so I, I think I touched upon this a little bit when I, you know, the last time I, I, I spoke about it, but I think ultimately it's kind of this. Ability to give people multiple options and tools to put into their toolbox so that they can kind of. You know, address their issues on a kind of basis, right?

[00:40:35] So if, you know, your lake has have issues and, you know, you're, you're looking into them, you may be more interested in what is causing those have issues rather than just, you know, just regular monitoring. Right?  so, you know. That's 1 thing. And the more options we give people, you know, the better it is.

[00:40:56] So the more technologies we can leverage to answer those sorts of questions,  you know, the better we get at kind of understanding this, this crazy world of haves around us. Right? And, you know, the, the, the real key here is getting all of these instruments and all of these data sets to talk to each other,  having a repository where,  Somebody can pick these things off up off of a shelf and,  you know, use them.

[00:41:24] And, you know, 1 unnamed lake in the middle of Alaska can decide, you know, hey, we're going to get some funding. We're going to buy 3 different types of sensors and we want all of that information to talk to each other and go into this repository. So, ultimately, it's kind of like,  you know, Socializing and, you know, making these types of monitoring techniques more engaging for the average person that volunteers that a lot of these like associations, we work with a lot of groups that, you know, they're, you know, the biggest users and biggest contributors to their data sets are retirees that live on the sides of these lakes.

[00:42:08] Right?,because they have all day and they look at the lake all day and they can note. All of these tiny little observations and they pick up on a lot of things, right?  so, so the closer we get to basically making it as, you know, as simple as we can to collect really, really high quality information,the closer we get to that you know, holy grail of all of this data going into, into one repository where everybody can kind of access it and you know, use it for high end science and, and answer all the tough questions.

[00:42:44] And then once we start answering the tough questions with monitoring, then we can start digging into what that actually means in terms of things like gene expression,  toxins, all of that kind of stuff. And then eventually mitigation and hopefully solutions.

[00:43:00] Ryan Sorichetti: Very interesting. Yes.  so, Chris, it sounds like as you're developing and optimizing these buoys and those networks for strategizing around where and when and what you're detecting, what's your long term vision?

[00:43:14] What do you see in terms of the practical application of these systems? Who's using them? And,  some of the long term benefits that we can see or expect in the future.

[00:43:27] Chris Lee: Well, I mean, as far as who's using them it's anyone from drinking water utilities or the, the, the,big 1 for sure.  states counties that have that are responsible for waterways. It could be the USGS. It could be others. Of course, lake associations.  yeah, that's kind of in order of. Of willingness and ability to,  kind of spend money on on the problems, I think, and being able to address them directly.

[00:43:59] So, I think,  the more you can, as I think we're all kind of converging on the more you can combine these,I think part of it will be combining the data,  without having to combine hardware.,but I do see the ability to have individual hardware things,objects doing more,  pretty much anything that,can be built.

[00:44:26]  you know,for a million dollars,eventually can be built for 110.  so it's, it's a matter of time for a lot of these things.  you start looking at things like micro fluidics, and you've got problems associated with that file and whatnot, but, um. I think there's a lot of solutions and the main thing is,  even before you get those hardware improvements.

[00:44:57] Let's try to get answers to the, to the burning questions, which the burning questions isn't tell me the details of what I have. The burning question is. Am I okay now, you know, was I wasn't I didn't like what I saw yesterday. Can I get to where I'm happy with what I'm seeing today?  and quite often that can have a black box aspect to it.

[00:45:21] We don't really know what  we're doing. Exactly, but we know that these things help them this  you know, we're in a better situation now.  so, yeah, we'll keep striving for the perfect,items, you know, sensors and stuff like that. But I think we can get a better,big picture as well. We're looking at 1 of the things that we're developing,with with NOA is the ability to distinguish classes on a continuous institute device.

[00:45:54] Classes of algae, like the five major group, or, you know, maybe, maybe some more discrimination than that. And so it's a development in progress. We don't know the answer as to whether we'll get there or not, but,there's a lot of activity happening.  so I think it, it, and again, I'm gonna,emphasize that it's happening on the mitigation side and the treatment side too.

[00:46:17] I'm really excited about some of the stuff that's happening there and,  working with those teams as well.

[00:46:26] Ryan Sorichetti: Excellent. Thank you for that.  so we are having a number of questions arrive in the chat here. And before I turn it over to those, I was just wondering if any of the panelists here had any final thoughts or something that they wanted to mention before we move over to that. And encourage those who have questions to continue putting them in the chat.

[00:46:51] Okay, so opening up the chat here, I do see a question for John. So, John, do you know of any cases where this wealth of understanding into toxic gene regulation and expression has been or can be applied to modeling and prediction tools?

[00:47:07] John Higley:  yes and no. So,yes, we've done this kind of modeling. I can't talk about some of the specifics.

[00:47:15] One of the issues is that we're the only private company that does environmental RNA,  you know, we, from a lab perspective, we always talk about DNA protein, right? You know, that's your basic bio 301 DNA is the good, the bad and the ugly, right? DNA is good. Super easy to work with. It's very stable.  RNA, you look at it funny, it falls apart.

[00:47:37] You basically have to turn into DNA work with it. And proteins are just ugly. There's too many of them. They're very complicated. They all bind differently and what not. So when we talk about like,tools like, like the, so if you want to look at toxin expression or the amount of toxin, the DNA doesn't tell you.

[00:47:54] It's a big family of genes. A lot of different species have kind of the same family. What matters is what is actually expressed. So that's where the RNA comes in. This is a great proxy for how much. Toxins actually expressed because this is a signal saying, Hey, produce this toxin and we can get very specific with it and we can get very quantitative with it.

[00:48:15] Super helpful. Proteins. That's the actual toxin. What is expressed. So we look at the proteins and you can do an ELISA. There's some issues with that because again, proteins are complicated your and there are 50 of these micro system genes. These micro system toxins. They're very, very closely related. So the specificity of the ELISA is not perfect, but They're pretty darn good.

[00:48:41]  and ultimately you look at mass spec where you're pulling them apart by amino acid by amino acid to exactly say, this is what this toxin was. So these are kind of the suite of tools, which is general overview. So we talk about where has this diversity of expression been utilized. So, in Austin,  they never had a, Lake Travis and Lake Austin never had a, toxic algae bloom,  for any, you know, in no memory, until there was a zebra mussel infestation, it changed the ecology of the lake pretty significantly, all of a sudden now they're getting toxic algae blooms every year.

[00:49:18] So, there was a lot of analysis of what those were, how much toxin they were producing, where were they, figured out focus areas for that. So, and that was utilized to predict or to,to,strategize where to put focus. And then a few years ago, it was,2021, 2022. there's that big, crazy ice storm that came down to Texas that had never really happened before.

[00:49:44] Thank you climate change and all of a sudden you have this massive freeze effect. It was in February. I was living there at the time. And,  then all of a sudden, there was a very unexpected  toxin bloom. You know, we had some dogs die and things like that, to drink the water, and it was completely unexpected.

[00:50:02] This is the middle of February. But what happened? Like Tom was saying, these big, Cyanobacteria now are broken apart and it was an antitoxin A and then a subunit of. And that was another thing that now all of a sudden we got to look at antitoxin production throughout the winter and summer months to predict the severity of these kind of things.

[00:50:21] We did another project with a municipality. I can't get into the details of this because it's they want to keep it tight to the close to the vest and that's the client's prerogative. So I have to respect that.  but this was another one kind of looking at. Nutrient loading,  where the nutrient basically looking at where,  you know, kind of the,the, the, what was causing the blooms, right?

[00:50:47] What was the nutrient loading coming from specifically around like,fecal contamination and then analyzing that collating that with, some public health. Issue bacteria, co relating it with cyanobacteriand co relating it with taste and odor bacteria families, and then determining kind of the root of that problem.

[00:51:08] That wasn't a prediction. That was like a, what do we tell these people? But ultimately, these tools have to be implemented over a Long period of time to get to where you can really create a full predictive model. And that has not occurred yet.  and hopefully it will eventually, but it's, it's, it's not there fully.

[00:51:26] So we have some examples where it's been implemented with some success for. Mitigation purposes, we have good knowledge that it works the way that it's supposed to work, but the full implementation of where this could go, or should go has not occurred.

[00:51:45] Ryan Sorichetti: Okay, great. Thank you. And I think that addresses 1 of the questions also in the chat about understanding what triggers toxin production and pointing that since it's at the molecular level, whether transcript, transcript techniques are part of the answer and the possibilities that exist.

[00:52:00] John Higley: Oh, yeah, there's like 12 operons. That are all impacted by different environmental factors. There's 3 that go 5 to 3 and the other ones go the other direction and which genes those are turning on that varies by species and ecological impact factors. And there's, that's an area where we have some good information that we need to do some more research on, but that's on the academic side, not on my side.

[00:52:23] Ryan Sorichetti: Okay, thank you for that. I'll pose a broader question to the group. No one in particular, if you have some experience or ideas on this, it'd be great for you to weigh in. Is there any information to help farmers deal with drought conditions, harmful algal blooms and build dugouts on the farms to grow crops?

[00:52:41] And are there impacts on nearby surface waters and rivers, creeks, et cetera?

[00:52:52] Thomas Bridgeman:That's a tough question. You're asking a bunch of algae people about farming, but,  there, there can be impacts, you know, especially,if there, if there's a large rain events, if, if farmers are using manure and there's large rain events, then, and the manure gets washed into nearby ditches and streams that can.

[00:53:17],create anoxic conditions that then kill the,you know, kill the fish and organisms  in a stream. So,manure runoff is definitely a problem.  there are the question about dugouts that meaning, meaning retaining water on the farm. Maybe that's something that is being implemented.  in the state of Ohio, there's a big program to create wetlands to hold water runoff running off from farms and process that water.

[00:53:46] So that, you know, muddy nutrient laden water goes in 1 end and clear water comes out the other end of these wetlands. So that's a big program.  there are. Programs for farmers to be able to turn off their,tiles in in this part of the country. There's a lot of tile drains under farms land so that they drain, but those can be turned off at.

[00:54:09] Sometimes they don't always need to be on because you don't always need to drain the water off the land. So there's a lot of different agricultural things. A lot of research in this area to buy agricultural academics.

[00:54:22] Ryan Sorichetti: Great, thank you, Tom.  there's a specific question for Igor.  how do your clients decide when a sample indicates a bloom?

[00:54:28] So, I guess what defines a bloom in that perspective and how are people deciding from the tool that you have what constitutes a concern or an alert for more action?

[00:54:37] Igor Mrjden: Yeah, so the annoying part of this is that, um. Each state, each. Municipality each country has their own criteria for what is a bloom or what data they're going to use to judge whether something is a bloom or not.

[00:54:56] So, what we do is we basically use the size, shape and classification of each colony that we identify using our  machine learning model,  the AI to basically estimate the cell count on a per genus level.,and so basically what you can,what you get with your report is this is how many colonies we detected from this genus.

[00:55:18] And this is the approximate cell count that the model,  has extrapolated out on a cells per milliliter.,level, we're also working on incorporating bio volume,into our measurements, which is basically just a geometric representation of the total volume taken up by a sign of bacteria, which is something that,  some places use.

[00:55:42],so ultimately, it's a question of how do you want to work with your local health department? The best way that we found so far is basically you run the sample yourself, you get the results and then you forward those results to your local health department. And you say, this is what we found today.  should I be taking any action?

[00:56:03] And because of that, we basically flag every  sample submitted that even has 1 sign of bacterial colony,as a.,you know, have tag that basically says this requires your attention. Please look over these results, forward them to your to your local health department and so forth. So, the key there is, you know, giving the clients the data they need in terms of cell counts, but also working closely with your local health departments and your local public health authorities to make those decisions.

[00:56:35] Yeah.

[00:56:36] Ryan Sorichetti: Okay. Thanks very much Igor. There's one more here that I think if someone has an idea on how to answer relatively quickly, I could send the rest through email to the panelists to see if anyone's able to answer. But the last question is modelers want to ground truth their model HAB results. So, for example, the biomass using real observational data.

[00:56:54] Often people are using satellites as the source of real data. Is this acceptable? And if not, what would be the best or most suitable method to do these comparisons? And again, this isn't at anyone specific.

[00:57:06] Thomas Bridgeman: So, yeah, Tom, go ahead. I can just I'll just say 1 thing real quick on that. If you talk to somebody people who do remote sensing, they will say that ground truthing is not ground truthing, you know, because you're taking a sample from a body of water.

[00:57:22] You're just grabbing a sample of water and the water that's. 10 meters away could be different. So you are, you know, your sub sample that you're taking may not be exactly representative of ground truthing. So ground truthing is kind of a subjective term, whereas the satellites are integrating over pixels that may be what 300 meters square or something like that.

[00:57:43] So which one is better kind of depends on what you're looking for. But I wouldn't say the ground truth is always the absolute truth.

[00:57:52] John Higley: Yeah, this is 100 percent agree. This is the issue of when I was saying about zooming out and zooming in, right? You know, so what does it really mean? It depends on what the answer is that you're trying to get to.

[00:58:05] And there are again, From satellite imagery is great. They're getting better. There's aerial photography. There's in water, you know, there's fluorometry. There's ELISA there's molecular biology. There's so many great tools that are at our disposal. They'll tell you different things. Taking a water sample at an individual location does not tell you that much.

[00:58:25] Unfortunately.

[00:58:27] Igor Mrjden: Yeah, and just to piggyback off of that,  you know, we, we started bloom optics as, you know, and  when we  first started, we were looking into using drone imagery,and hyperspectral multispectral imagery to take aerial photography and map blooms and things like that. But,you know.

[00:58:45] You know,Tom and John are 100 percent right. It's hard to say, oh, you know, this 300 meter pixel represents this 1 data point. But the nice thing about ground truthing,is that it's getting better in that, you know,  with like instruments, like with Chris's buoys or our instrument. You can cover more area and get more representative data.

[00:59:10],to then maybe it's not ground truthing, but it gives you more apt comparisons between the 2 technologies right?  you know, with our technology for for the price of testing, you know, 10 to 20 samples in a laboratory, you can test on limited samples during the course of the summer.,so you can collect 100 samples instead of instead of 10 every day and run them through the system and,  you know, that.

[00:59:38] All of a sudden becomes a little bit more representative of the total body of the lake. Right? Of course, that gets into the issue of getting out to those sample locations to collect the samples and everything like that. But,you know, we're, we're, we're getting better and, you know, satellites are going to be improving,  you know, quite rapidly, hopefully, and those imaging techniques, those, those pixels are going to get smaller.

[00:59:59] Hopefully, and,  they're going to get more representative on, you know, more localized locations. So, you know, those technologies are very much improving. But right now, as it is, like, John and Tom said, it's not a 1 to 1 relationship.

[01:00:14] Ryan Sorichetti: Excellent. Okay. First, I'd like to thank the panel experts very much for your time today, sharing your knowledge, experience, insights.

[01:00:22] I'd like to thank everyone for joining. Again, please tune into the final webinar in this series that's hosted by the Cleveland Water Alliance and the Great Lakes Habs Collaborative. Again, it's on Thursday, October 17th, and the topic of that one will be innovations in nutrient monitoring technology. So have a great afternoon and thank you all very much for your time today.